skip to main content

Model-based Dose Escalation — SCE Medical Oncology MCQ

Instant feedback + full explanation. One question, done properly.

ModerateClinical Trials & StatisticsModel-based Dose EscalationSCE Medical Oncology

A 55-year-old woman with cancer is referred for a phase I trial. She asks about the difference between the traditional 3+3 design and model-based designs (e.g. CRM/BLRM). What is the advantage of model-based dose-escalation?

Educational content. Not a substitute for clinical judgement or local policy.

Reveal the answer and explanation

Correct answer: B — Model-based designs (CRM, BOIN, BLRM) use statistical models updated with real-time toxicity data to recommend the next dose — they are more efficient (treat more patients near the MTD), more accurate at identifying the true MTD and expose fewer patients to subtherapeutic doses compared with the rigid 3+3 algorithm

Traditional 3+3 design is simple but inefficient: it treats many patients at subtherapeutic doses, is poor at identifying the true MTD (correct identification ~30-40%), and is rigid. Model-based designs (CRM — continual reassessment method; BOIN — Bayesian optimal interval; BLRM — Bayesian logistic regression model) use accumulating toxicity data to continuously update dose-toxicity models, more efficiently guiding dose escalation. They treat more patients near the MTD (~50-60% vs ~25-30% with 3+3) and more accurately identify the true MTD (~60-70%).

Reference: JRCPTB Curriculum; Le Tourneau et al JCO 2009; Yuan et al Clin Cancer Res 2016